Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

I guess it's interesting that these studies fail to replicate. But I fail to see how saying a "lack of theory of mind" is not a very accurate way to communicate the experience and symptoms of autism.

Autistic people, by their own account and by well documented empirical evidence, express and demonstrate challenges with understanding sarcasm, irony, nonverbal cues, idioms, figures of speech, etc. Autistic people have a unique set of difficulties with maintaining friendships and employment.

Now, I'll take two definitions of Theory of Mind:

> Theory of Mind is the branch of cognitive science that investigates how we ascribe mental states to other persons and how we use the states to explain and predict the actions of those other persons. [0]

> In psychology, theory of mind refers to the capacity to understand other people by ascribing mental states to them. A theory of mind includes the knowledge that others' beliefs, desires, intentions, emotions, and thoughts may be different from one's own.[1] Possessing a functional theory of mind is crucial for success in everyday human social interactions. [1]

So... I'm confused. If autism is the impaired ability to ascribe mental states , how then is autism not an impaired ability to form theory of mind? Are we just arguing semantics here?

The authors of this research might be resting their case on a slippery re-definition of theory of mind?

> The assertion that autistic people lack a theory of mind—that they fail to understand that other people have a mind or that they themselves have a mind—pervades psychology.

Of course Autistic people understand that other people "have a mind" or "that they themselves have a mind".. This seems ridiculous. I have had many relationships with autistic people and obviously they know what a mind is and that other people have a mind.

[0] - https://iep.utm.edu/theomind/

[1] https://en.wikipedia.org/wiki/Theory_of_mind



I have a number of thoughts on this.

> If autism is the impaired ability to ascribe mental states

First, autism is not the impaired ability to ascribe mental states. Earlier in your post you more accurately described that as a symptom.

There are attempted theories that get more to what autism "is", that it's a heavier weight on bottom-up processing, processing distinct details more and holistic things less. Another (congruent) idea is related to a "narrow tunnel of attention" and to using more of the brain to attend to a more narrow set of stimuli and thoughts. (You can look up monotropism for more info on this idea.)

Secondly, in my experience and others I've read - autistic individuals seem to often do much better at understanding each other. I mean this in two ways: autists are often better at understanding other autists than they are at understanding neurotypicals, but also in many cases autists may be better at understanding other autists than neurotypicals are at understanding autists.

With this framing you could just as well say neurotypicals lack a theory of mind. Largely we don't do that because there are more neurotypicals than autists.

Of course that's not the full reason and it's not fully symmetrical. If there were a 50/50 split in the population then I think autism would be less of a disability for folks in terms of this theory of mind stuff, but it would still be a significant disability for anyone who struggles with all the other consequences of autism.

But it is still relevant that this "theory of mind" stuff may be a misleading way to talk about it; by default I would expect everyone to have an easier time with a theory of minds like their own. And thinking about it this way also helps inform neurotypicals with autistic loved ones that if they expect their loved one to work hard to understand them, they could do the same.


"Secondly, in my experience and others I've read - autistic individuals seem to often do much better at understanding each other. I mean this in two ways: autists are often better at understanding other autists than they are at understanding neurotypicals, but also in many cases autists may be better at understanding other autists than neurotypicals are at understanding autists."

This is an interesting point.

"With this framing you could just as well say neurotypicals lack a theory of mind. Largely we don't do that because there are more neurotypicals than autists."

I think the really interesting result would result from testing this.

i.e. because of prior knowledge, neurotypicals model the default person as neurotypical.

However, knowing whether the second party is neurotypical or autistic, can a neurotypical person model their second party's mental states better than an autist, or worse?


I also found that post very interesting. I would also make a point (vapid as it may be) that autism is a spectrum. Many people over the years have called me autistic, and I for sure have a few too many autistic family members for it to be a fluke. Somehow though, in person I have a solid skill at teaching people technical information at a level they can understand it, I am quick to identify what types of analogies they will understand and my technical knowledge is generally broad enough to relate it. If you saw me at work you would probably consider me -a hard worker -a dirtbag -adept -foolish, whatever it is, but likely not autistic; if you got to know me, you would probably be suspicious. 2 sibblings are in a similar category but more advanced autism, 1 undiagnosed, they are perfectly capable of reasoning about their own/others cognition.

It makes me suspicious though, if I was more defensive then I wouldnt go around saying I might be mildly autistic. If that other me were surveyed, would I unwittingly bias the results?


> Somehow though, in person I have a solid skill at teaching people technical information at a level they can understand it, I am quick to identify what types of analogies they will understand

That's interesting. There's something odd about me but I don't know what, I'm not like other people and they tend to consider me a bit weird (usually in a nice way). In turn, I often can't understand them. That said, what I just quoted above of yours pretty much describes me. I'm extremely good at putting myself in another person's boots for teaching technical subjects, and indeed I'm not too shabby at the social side of things - I'm often surprised at people's faux pas. And yet, people's behaviour regularly baffles me too. It's like I'm picking up a lot of some things but perhaps too little of another.

I did an online test for autism, expected to come out below average but much to my surprise came out above.

> you would probably consider me -a hard worker -a dirtbag -adept -foolish

Looks a bit like you're passing CLI arguments :-)


> I would also make a point (vapid as it may be) that autism is a spectrum

I don't think that's a vapid point at all. It is however a widely misunderstood statement. Autism being a spectrum does not mean "you can have it a little, or a lot, or in-between". It means you can have any one of the symptoms a little or a lot or in between, individually.

Among other things this means being good socially does not trivially preclude autism.


> However, knowing whether the second party is neurotypical or autistic, can a neurotypical person model their second party's mental states better than an autist, or worse?

Even this is still limited by the neurotypical person's prior experience with autistic people and recognizing them as such.

All of this really reduces down to "it's hard to understand people whose brains work different from yours or from others' that you have significant experience with."


>in many cases autists may be better at understanding other autists than neurotypicals are at understanding autists. With this framing you could just as well say neurotypicals lack a theory of mind

Excellent comment, which has me rethinking autism. It's a topic I've thought quite a bit about, since by all measures I am the polar opposite of autistic, I'm extremely abstract to the total neglect of detail, and yet I've always gravitated to friends high in autistic traits. Maybe they have or see something I lack.

The DSM-V defines autism by the disability it causes. The diagnostic criteria list are all things neurotypicals can do but autists struggle at. So I always saw it framed as a disability.

But could an equally compelling list of diagnostic criteria be written, which is all things autists can do but neurotypicals struggle at? Could neurotypical be framed as a disability?


The DSM-V is about clinical diagnoses. Mental disorders are understood as something everyone experiences to varying degrees. You're considered to have something clinically when it impairs your daily life, something psychiatrists have defined as being 3 σ away from the average number of symptoms someone would report.

In this example struggling with things like math wouldn't qualify as a disability unless it impaired you in a statistically significant way


> But could an equally compelling list of diagnostic criteria be written, which is all things autists can do but neurotypicals struggle at? Could neurotypical be framed as a disability?

Something like this? https://lemonandlively.com/allism/

From my perspective, the surprising thing about people not on the autistic spectrum is how often they lie, how often they stab each other in the back for quite trivial reasons, and how obsessed they are with evaluating and maintaining everyone's position in the pecking order -- they are barely able to think about anything else, at least half of their brains are always evaluating this.

And the famous "theory of mind" in most cases is merely the ability to understand someone who is almost the same. I mean, of course if someone thinks exactly the same way I do, it will be very easy for me to predict what they think. But find someone different (doesn't need to be autistic; any kind of difference will work) and suddenly most people's ability to read their mind is gone.


At a high enough level; science, math, software; anything that requires super human focus basically.

Speaking for myself, I had to learn everything about human behavior from the ground up, which paradoxically makes me very good at reading and dealing with other humans IRL since it's all conscious for me.

But, it's super important to remember that it's a spectrum, some people are seemingly seriously disabled by it.


> Could neurotypical be framed as a disability?

No. What is it with so many autists trying to turn things around and act as though they are better in some way and NTs are the 'problem'. It's bizarre and seems like some sort of denial.


This is just nonsense. Depending on the severity, autism makes a person completely dependent on other normal people (neurotypical is just a pc term) to simply survive. There's no degree of "normalcy" that, the more "severe" it gets, the more dependent on others you become.


That is true, I was ignoring the severely autistic. It becomes crippling at a certain point, but on the other hand, there are billionaire CEOs on the autism spectrum, who probably got there because they are different.


I'm still not convinced were not arguing semantics. Okay, agreed, an impaired theory of mind is a symptom of autism. But to me this is like saying the sky is blue, and then someone saying "no its actually scattering blue".

The fundamental mechanisms that make an autistic brain different than the statistical average brain are not yet fully understood, I know. But this study we are talking about is not about the fundamental mechanisms.

I just got out of a relationship with a French woman because of our differences in cultural norms, she also had an accent and we used idioms and expressions that neither of us were super keen on. We both lacked a robust theory of mind of each other. I had an impaired ability to ascribe mental states to her. And I find it very unlikely that I will ever become fluent in French. I currently and probably always will lack that capacity.

I also have an impaired ability to ascribe mental states to Donald Trump, and my cat. But most people would consider me allistic. Because I have mental processes that allow me to ascribe the mental states of other statistically normal people. Within the context of statistically average people, of my own culture, I have a robust-enough theory of mind. ToM is not binary, but a degree, and context dependent.

"Mental conditions" are classified mostly because they are disadvantageous or uncomfortable to some group of people. They can only be classified in relation to normal social behavior at some degree of deviation. Shouldn't it be implied that allistic people also have an impaired theory of mind of autistic people.

Taking Wikipedia's definition of ToM, it specifically mentions "capacity". So when you mention that autism may be a "narrow tunnel of attention", this feels very consistent with the definition of ToM. If my attention is extremely narrow, then I will have an impaired ability to understand other people by ascribing mental states to them.


> So when you mention that autism may be a "narrow tunnel of attention", this feels very consistent with the definition of ToM. If my attention is extremely narrow, then I will have an impaired ability to understand other people by ascribing mental states to them.

Indeed that may be a cause of impaired theory of mind in many autistic individuals. It would depend on what the narrow tunnel of attention is focused on. A narrow (and deep - I failed to say that earlier) tunnel of attention focused on another mind, say an autistic psychotherapist, or better an autistic psychotherapist focused on autism, may have a greatly heightened theory of mind.

Perhaps tangential, but consider Temple Grandin's exemplary work designing animal handling facilities that work much better largely due to treating the animals better - often with changes or accommodations to things that nobody realized caused the animals stress and anxiety, often by putting herself in their place.


Condensing GPs wall of text, I think the overall point is that Aspies are bad at ascribing motivation and mental states to neurotypicals, but much better than neurotypicals at ascribing them to other Aspies.

It's more a case of speaking different "emotional languages" than a strict deficit.


> With this framing you could just as well say neurotypicals lack a theory of mind. Largely we don't do that because there are more neurotypicals than autists.

It's because NTs are the default, and Autism is seen as a (sometimes) harmful deviation that is still being investigated.


I think they are claiming the opposite and that autistic people don't lack theory of mind, but that the theory of mind claim is stemmed from their communication impaiarment. I am no expert, but I think they are trying to overcome the idea you're presenting.

> Because theory-of-mind tasks rely heavily on “fairly complex language” (San José Cáceres, Keren, Booth, & Happé, 2014, p. 608) and because autism, by diagnostic definition, involves communication impairment (Gernsbacher, Morson, & Grace, 2016), it is unsurprising that autistic participants with communication impairment perform less well than nonautistic participants without communication impairment. And because autistic people vary in their communication impairment (Gernsbacher, Geye, & Ellis Weismer, 2005), it is unsurprising that autistic people vary in their theory-of-mind task performance.


This was my takeaway as well - theory of mind can sometimes be described as 1. "literally does not understand that other people have their own thoughts" or 2. "cannot reliably predict what other minds would think". Most high-functioning autistic people I've met clearly don't meet the first definition, but definitely meet the second.

The second definition makes more sense to me since it also comes up in non-autism diagnoses, like personality disorders. They have trouble reliably predicting what other minds would think, but for different reasons.


High-functioning autistic people would meet the 2nd definition when trying to predict what neurotypical people would think. But they're also probably better at predicting what other high-functioning autists would think than neurotypical people are. It's just easier to imagine what someone might think if they think like you.


A part of all psychology diagnoses is that it must both match a definition and negatively impact someone's life. So if you never had to care about what neurotypical people think, then it wouldn't rise to the level of a problem and there's no diagnosis.

You see people care about this stuff because they go to their therapist and say, "boy there are a lot of neurotypical people I have to predict, and being slightly better at reading autistic people doesn't make up for it."


Unreliability is relative. Everyone botches this up all the time. The most neurotypical person I ever met literally questioned the whole premise of predicting what other people were thinking and said that most people will just tell you things if you ask.


> most people will just tell you things if you ask

In my experience people are often bad at explaining their inner emotions. Rationalisation is one of my favorite words!

I will often ask a friend why they acted a certain way, and I will carefully listen to what they say, but I often disagree with their stated motives - denial and lack of self-awareness are rife. This has nothing to do with people consciously trying to deceive me (there's another topic!). I have the same faults. I'm average at reading people, yet all too often I have to trust my own flawed instincts about what I think is going on.


> Autistic people, by their own account and by well documented empirical evidence, express and demonstrate challenges with understanding sarcasm, irony, nonverbal cues, idioms, figures of speech, etc.

Last year I was officially diagnosed with Autism Spectrum Disorder (ASD) Level 1, which closely aligns with what was once categorized as Asperger's Syndrome.

What you're saying is a generalization I would've made at one point, and may very well be true of other people with ASD. It's generally not true for me, and in fact, I feel like I have a preternatural sense for what I think linguists call "pragmatics". I grant that I may have had to reverse-engineer my way to that understanding, since I don't know how more neurotypical people experience them.


I think this is all utterly misunderstood. IMO autism is a predictive processing failure: the brain gives too much weight to inputs and inference and not enough to prior knowledge. This causes hypersensitivity to everything surprising, which is deeply unpleasant. The coping mechanisms are to try to shut out anything that's insufficiently predictable, to try to force things to be as predictable as possible, or to restrict attention to controllable and predictable things.


Regarding the fundamental mechanisms. This is the view that I subscribe to as well.

It explains "stimming". Repetitive behavior is soothing because its predictable. And it explains things like toe walking - feet are sensitive, and the inability to predict the stimulus of things we might step on is uncomfortable. Folks with autism may have a heightened discomfort with unpredictable sensations.

There is growing evidence that atypical dendrite development in childhood and early adulthood is linked to autism, schizophrenia and epilepsy.

I see autism and schizophrenia on a spectrum of predictive processing failure. Normal folks typically cannot tickle themselves. But folks with schizophrenia can - probably because its difficult for them to predict their internal sensations.


[I'm somewhere (though I don't think very far? on the spectrum).] I think the way we go about diagnosing or assigning value to abilities is productive. I for one can if I really set my heart/mind on it, make a good guess why someone I know did something. I rarely do this though--it seems like such a waste of time, when I could just ask them. I also can't stand gossip which is a collective theory of mind.

So I don't see the need for the part "and how we use the states to explain and predict the actions of those other persons." As for the 2nd definition other than finding oneself in awkward situations, e.g. being the last one still at a 'party' not realizing that the host wanted you to leave an hour ago, I haven't really found it to hinder my success. I also don't think it's one thing, black/white, but if I did have to pick one I'd say that the person somehow naturally places their attention on something other than people around them almost all the time starting from a very early age.


>express and demonstrate challenges with understanding sarcasm, irony, nonverbal cues, idioms, figures of speech, etc.

A common element to these, is that they are either a bit sideways of a straightforward relationship or many times are being used directly to deceive. Basically, my impression is that autistic people and myself compared to others are by default worse at detecting “liars”.

Most animals on the other hand, don’t try to deceive or control their communication methods in order to hide what they are really thinking. When they are happy, sad, or angry, they consistently display exactly those things. This is why I find it easy to emphasize with animals who have a toddler level of intelligence.


> Most animals on the other hand, don’t try to deceive or control their communication methods in order to hide what they are really thinking.

I'd challenge that assertion. It may be literally true if you allow "all animals" to be skewed by masses of beetles, but then it doesn't seem very relevant to the subject of human cognition.

Conversely, animals that are related to us--or which have traits we consider intelligent--are often capable of deceiving others of the same species.

Probably far more than we know of, given that to detect their deception, we'd first need to understand animal communication. Consider cuttlefish that communicate by skin coloration. They may lie by simultaneously sending contradictory messages on each side of their body towards different recipients.


Chimpanzees engage in conspiracy and betrayal as a common feature in their societies. Domesticated animals may show humans what humans expect to see, that's all.


> Domesticated animals may show humans what humans expect to see, that's all.

On that note, It's been shown that in a lab--the location, not the breed--that dogs can use deception to get treats. (Not a surprise to most dog-owners.)


Not to mention this presumes some "default" way the mind works that is commonly presumed and commonly accurate, which seems transparently absurd.


99% of people put the definitions on 1% of people, on how well the 1% interface with the 99%. What if it was the other way around? There is a neurological difference, but is it really neurologically worse? Nature decides I guess. Well it would, but these days society decides, well, the bulk of society, well the 99% meaning nonautistics that is.

My personal problem is that, there is some pervasive belief that autistics lack empathy, morals, or gets judged as being narcissistic or worse. These sorts of accidents can happen because signalling doesn't happen, because of the neurological difference, despite the autistic person in question hypothetically following a very ethical code and behaves with full integrity and love and so on but not initially obvious due to the neurological gap. My fear is that shaky concepts, like ascribing lack of theory of mind, could contribute to "othering" and even more misunderstanding.

So in some ways it becomes a semantic argument, theory of mind, why yes minds exist, and autistics may be bad at ascribing mental states to allistic people, but allistic people are bad at ascribing mental states to autistic people (guess you'll have to take this one for granted, but e g or they could replicate the specialised results I mean, like systematise vast bodies of knowledge) so I would say it goes both ways right? So what I mean is the autistics could say allistics lack a theory of mind, because they only understand eachother and not autistics. Except the weight is put on the 1%, by the 99%, rather than a uniting force for humans to understand eachother, it becomes a onesided label of weakness.

edit: I wrote this related comment in another thread as well, explaining why I posted this link : https://news.ycombinator.com/item?id=39253047


I'm confused. Elon Musk says he is autistic (actually he says he has aspergers, which has been eliminated with DSM 5 and replaced with autism) , is there any evidence a high functioning autistic person like him "express and demonstrate challenges with understanding sarcasm, irony, nonverbal cues, idioms, figures of speech, etc".

This definition- of someone failing to understand idioms or figures of speech, sounds more like Lieutenant Commander Data on Star Trek than the meme twitter billionaire.

Is the definition of autism so broad as to have little practical meaning?


It is certainly very broad. And as much as some people try, you can't make any complete generalisations about what autistic people can or cannot do.


What was described was a specific common experience, not a definition. There are strengths as well to autism. The definition is broad, but it does have practical meaning, diagnosis is done by educated psychologists obviously though




Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: